An ab initio study of the hydrogen-helium phase separation in the Gibbs ensemble

The giant planets observations lead to significant advances in our understanding of dense plasmas. The detection of a strong magnetic field around Jupiter was an indirect proof of hydrogen metallization at high pressure. The excess luminosity of Saturn as well as the gravitometric measurements by Cassini and Juno strongly support the idea that hydrogen and helium segregate at high pressure. Ab initio simulations confirmed these predictions although a large uncertainty remains regarding the critical surface of the H-He phase separation. Recent experiments were able to detect a phase separation of hydrogen-helium in the megabar regime using reflectivity measurements in laser-shocks. However these experiments place the critical surface of the H-He demixing at much higher temperature than the ab initio studies.

In order to alleviate some of the numerical simulations limitations and uncertainties in exploring phase separation, the Gibbs ensemble is a very powerful statistical ensemble. It consists in simulating, within a Monte-Carlo framework, two independent boxes that are allowed to exchange volume and particles. We combined this ensemble with a density functional theory code to consistently explore the phase diagram of hydrogen-helium. In this presentation, we will show how to perform robust ab initio Gibbs ensemble simulations for dense plasmas and discuss the specific results obtained for the hydrogen-helium demixing